boulos

82 ms·

841.
by boulos 6y ago
Hmm. I’m a big fan of immediate mode myself, because it’s kind of amazing how much you can actually push through via brute force. However, wouldn’t you agree that Unity, Open Scene Graph, etc. are successful counter arguments? I would say i
842.
by boulos 6y ago
I assume Raph meant in the screen reader sense of the term.
843.
by boulos 6y ago
I can’t seem to find the PDF of Brent’s siggraph sketch in 2006 (Shadow map bias cones), but our follow up work [1] hopefully makes the connection between distance and r^2 clear: these techniques are all computing some sort of solid angle a
844.
by boulos 6y ago
Public funding. If it’s a collective good, “we” should all be willing to contribute towards it.
845.
by boulos 6y ago
Not that it matters, but #6: https://kskedlaya.org/putnam-archive/putnam1999results.html
846.
by boulos 6y ago
Disclosure: I work on Google Cloud. The link here is just to someone’s personal Colab notebook (which Google hosts for free at colab.research.google.com). This doesn’t appear to be a Google project. It’s more like a link to a GitHub repo.
847.
by boulos 6y ago
They mention this in the writeup: it’s not going to be made public (yet). Also BuzzFeed News does real journalism, despite the BuzzFeed origin of their group.
848.
by boulos 6y ago
I was surprised that the tile-based updates from VNC, Teradici, PCoIP etc. wasn't in the related work. In particular, a challenge with "send < smaller > updates" rather than whole frames is how you decide to deal with m
849.
by boulos 6y ago
> Why is this stake necessary? Edit: it’s not necessary/a requirement. They’re noting that after the transaction, SoftBank will still fall under the 10% ownership threshold that requires more reporting from the SEC [1]: > Section
850.
by boulos 6y ago
I should clarify: A100s are likely a bad price/performance trade off for inference. The T4 part is designed for that, but doesn’t have NVLINK. Do you have models >16 GB that you’re trying to do real-time inference against? Feel fr
851.
by boulos 6y ago
Disclosure: I work on Google Cloud (but you don’t need to rent GPUs from us). Absolutely, but as last year’s discussion highlights, a bunch of GPUs connected via NVLINK kind of gives you the aggregate memory of the set for some of these dat
852.
by boulos 6y ago
My immediate question was “what’s their annual budget anyway?”. It seems like the lab is funded primarily from its initial endowment, and that for 2017 they only needed an extra $225k [1]. What’s unclear is whether the additional support fr
853.
by boulos 6y ago
> If there was a sports-like betting site for future decisions that major open source projects are going to make, I would put a thousand dollars on "Cargo/Crates will implement namespaces." There is, sort of! Assuming you
854.
Is it possible to “write” using speech-to-text software?(economist.com)
2 points by boulos 6y ago|0 comments
855.
by boulos 6y ago
Sure, but the parent comment was saying they're worried the Fed will lose all the money. But the Fed is buying things broadly (initially they bought Bond ETFs/funds!), so while some places will be higher, others will be lower. My
856.
by boulos 6y ago
Actually, your 25% delinquency rate seems unreasonable. During the financial crisis, the subprime delinquency rate peaked at 26% [1] but the overall (whole US market) rate was just over 9%. Delinquency is also mostly just about being a le
857.
by boulos 6y ago
Eh. This is sadly a clickbait headline with a murder-mystery writeup. The seismometer on the lander (sitting on Mars) registered a tiny tilt: > “But we didn't expect this seismometer reading; it's an unusual signal," Stähl
858.
by boulos 6y ago
Hmm. In the end, maybe transparency is a bad example, because people immediately jumped to approximating it instead. Wyman's review paper though is a good read regardless: http://cwyman.org/papers/hpg16_oitContinuu
859.
by boulos 6y ago
Raph, I think some of the order-independent transparency folks (Bavoil, Wyman, Lefohn, Salvi, etc.) have had to do this bounded k-stack plus overflow thing. Might be a good "hmm, how did they do it" for comparison.
860.
by boulos 6y ago
NYC issues student metro cards from the school if you need to go more than half a mile: https://www.schools.nyc.gov/school-life/transportation/bus-e...
861.
by boulos 6y ago
San Francisco has a means-tested variant instead [1] with the same thresholds for both children (5-18) and seniors (65+ here). Children under 5 are free. I think you can quibble about means testing, barriers to application, and just “do it
862.
by boulos 6y ago
The other option that Matt Levine has been writing about more lately is no longer having integer number of shares (jedberg’s precision question is similar). Amusingly, there are folks who take the other side of this argument, and very drast
863.
by boulos 6y ago
Hmm. Levels.fyi has a lot more data: https://www.levels.fyi/comp.html?track=Software%20Engineer&s... has just over 2000 entries for Software Engineers. Admittedly, it’s time since launching in 2017, but that’s still a
864.
by boulos 6y ago
As many have pointed out, this is a pretty shallow critique of a straw man wealth tax that nobody has actually proposed. If you want a little more depth on taxation proposals, I’d recommend the Saez and Zucman book [1] that underpinned most
865.
by boulos 6y ago
The original App Engine Cron had its own "human friendly" format for the schedule argument [1]. The problem is that what people want to copy/paste from the internet is some form of vixie-cron usually. However, all cron form
866.
by boulos 6y ago
I think that Coinbase’s Odin ( https://github.com/coinbase/odin ) is not the same as this person’s personal project. They’re certainly not the same repo.
867.
by boulos 6y ago
The normal first question is "do you have near-term need for this money" which it sounds like your answer would be "No, this came to me and now I just need to invest it for a long time". It sounds like your asset allocat
868.
by boulos 6y ago
Yes, (and I considered including that disclaimer). I give them shit about that all the time. I'm an equal-opportunity complainer.
869.
by boulos 6y ago
Jinx :). I really am surprised the marketing folks didn’t claim this was an ~3 exa-op supercomputer based on int4 w/ sparsity.
870.
by boulos 6y ago
Sigh, I wish people wouldn’t say “peta flop ” for these. https://www.nvidia.com/en-us/data-center/a100/ is the most official reference. If you scroll to the bottom, you’ll see that an A100 part can do ~20 Ter

More ›